An Automated High Aspect Ratio Mesher for Computational Fluid Dynamics

Computational fluid dynamics (CFD) simulations are routinely used while designing, analyzing, and optimizing air- and spacecraft. An important component of CFD simulations is mesh generation, or discretization into polygonal or polyhedral cells, of the domain being analyzed. The overall computational cost and accuracy of simulations depend heavily on mesh quality – the size, shape, and structure of the cells. Another important aspect of CFD simulation is that solutions are achieved iteratively, with each subsequent pass decreasing error and increasing solution accuracy. Grid adaption uses output from the last simulation to improve the mesh for the next. FUN3D is a CFD simulator developed at NASA that requires both a tight integration with mesh generation software for grid adaption and the generation of high aspect ratio cells (i.e. 10,000:1) to accurately capture dynamics around boundary layers. Current meshing methods use the well known advancing front or Delaunay algorithms, and the user must often perform multiple manual inputs and interactions to generate a mesh of sufficient quality. Ideally, a meshing module would subdivide complex designs and perform grid adaption automatically, with little or no human intervention. The proposed innovation is the development of a very high aspect ratio mesher that demonstrates a significant improvement over current techniques as measured by the time and effort necessary for FUN3D users to solve CFD problems.